Anthropic Releases Claude Opus 5 With Cost-Capability Toggle

Anthropic released Claude Opus 5 today, a new AI model designed for everyday business needs. A key feature of Opus 5 allows users to select the computational effort—low, medium, or high—the model expends on a task, enabling a balance between cost and capability. This addresses growing concerns from enterprise customers regarding the high expense of AI usage.

Anthropic stated that Opus 5 is more user-friendly than previous models, requiring less iterative prompting. The model is also designed to verify its own work and recover from errors autonomously. This release marks Anthropic's fourth model launch in less than two months, following Mythos 5, Fable 5, and Sonnet 5 in June.

Fable 5, one of the previously released models, faced export controls from the U.S. government after Amazon researchers reported bypassing its safeguards. Anthropic temporarily removed Fable 5 from the market on June 12, re-releasing it on June 30 with enhanced security. Business customers and developers had also criticized Fable 5 for its high token consumption, leading to significant billing costs.

Anthropic claims Claude Opus 5 offers capabilities close to Fable 5 but at half the price. However, the company still recommends Fable 5 for more complex, long-term autonomous projects. This focus on economical token usage mirrors OpenAI's marketing for its GPT-5.6 model, released on July 9. Anthropic is also navigating a balance in safety, aiming to be secure without being overly restrictive.